WO2020051635A1

WO2020051635A1 - 3d printing powder composition and a method of 3d printing an article

Info

Publication number: WO2020051635A1
Application number: PCT/AU2019/050971
Authority: WO
Inventors: Jay SANJAYAN; Behzad NEMATOLLAHI; Ming Xia
Original assignee: Swinburne University Of Technology
Priority date: 2018-09-10
Filing date: 2019-09-10
Publication date: 2020-03-19

Abstract

A powder composition for forming 3D printed article, the composition comprises: an aggregate; a cementitious material selected from Portland cement and/or Calcium Sulfoaluminate (CSA) cement to bind the aggregate when mixed with a binder to form a 3D printed article; and an accelerator selected from calcium aluminate, calcium sulfate hem i hydrate (plaster of Paris) or lithium salt to accelerate setting of the article formed; and a flow control agent to control flowability of the powder. There is also provided a method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving powder from the powder feeder, and a print head for applying a binder to the bed.

Description

3D PRINTING POWDER COMPOSITION AND A METHOD OF 3D PRINTING AN

ARTICLE

FIELD OF INVENTION

This invention relates to a powder composition for 3D printing, and a method of 3D printing an article, more particularly a 3D concrete printing (3DCP) method.

BACKGROUND ART

Three-dimensional (3D) printing, also known as additive manufacturing (AM) is a group of emerging techniques for fabricating a wide range of structures with complex geometries from digital models. The process involves printing successive layers of materials that are formed on top of each other. Several industries including aerospace, automotive, biomedical have already explored the benefits of adopting this technology as an integral part of their product manufacturing process.

In the recent years, the construction industry has attempted to adapt AM technology for construction applications owing to its potential use for“freeform” construction of buildings and other complex structures of virtually any shape without the use of expensive formwork. Elimination of the formwork would result in considerable cost savings, as the formwork represents 35% to 60% of the cost of construction of concrete structures. Besides, the elimination of formwork also reduces material wastage, and therefore improves sustainability in construction. The powder-based 3D Concrete Printing (3DCP) technique is a typical AM process that creates accurate structures with complex geometries by depositing binder liquid (or“ink”) selectively into to a powder bed to bind powder particles to each other. The printing process begins with a thin layer of powder (approximately 0.1 mm) being spread and smoothed by a roller over the powder bed surface. Subsequently, the binder solution is delivered from binder feeder to the print head and selectively jetted by the nozzle(s) on the powder layer, causing powder particles to bind to each other. Repeating the described steps, the built part is completed and removed after a particular drying time, and unbound powder is removed by using an air blower. This technique is typically an off-site process designed for manufacturing precast components and is highly suitable for small-scale building components such as panels, permanent formworks and interior structures that can be later assembled on site.

Although the powder-based 3DCP techniques can offer several advantages in the construction industry, there are a number of challenges to overcome before the technique can be fully utilized. One of the main challenges with the powder-based 3DCP technique is that proprietary printing materials that are typically used in commercially available powder-based 3D printers are not suitable for construction applications.

Therefore, there is a desire to formulate powder for 3D printing applications which are suitable for construction applications. SUMMARY OF THE DISCLOSURE

The present invention provides a powder composition that can be used for 3D Concrete Printing (3DCP) for construction applications.

In a first aspect, there is provided a powder composition for forming 3D printed article, the composition comprises: an aggregate; a cementitious material selected from Portland cement and/or Calcium Sulfoaluminate (CSA) cement to bind the aggregate when mixed with a binder to form a 3D printed article; and an accelerator selected from calcium aluminate, calcium sulfate hemihydrate (plaster of Paris) or lithium salt to accelerate setting of the article formed; and a flow control agent to control flowability of the powder. In one embodiment, the powder composition comprises Portland cement, amorphous calcium aluminate, hydrophilic fumed silica, and an aggregate. In a second embodiment, the power composition comprises Portland cement, calcium aluminate cement, calcium sulfate hemihydrate, hydrophilic fumed silica and an aggregate.

In a third embodiment, the powder composition comprises Portland cement, lithium salts, hydrophilic fumed silica and an aggregate.

In a fourth embodiment, the powder composition comprises rapid hardening Portland cement, hydrophilic fumed silica and an aggregate.

In a second aspect, there is provided a method of forming a powder composition for forming 3D printed article as previously described, the method including: mixing a cementitious material and an accelerator in a mixer to form a first mixture; mixing a flow control agent with the first mixture to form a second mixture; and mixing an aggregate with the second mixture to form the powder composition.

A third aspect of the invention provides a method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving a powder composition from the powder feeder, a print head for applying a binder to the bed, the method including: forming a layer of the powder composition as previously described on the bed using the powder feeder; applying binder to the layer of powder; drying the layer of powder for a predetermined time period; and repeating the forming, applying and drying steps until the article is formed. A fourth aspect of the invention provides a method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving powder from the powder feeder, a print head for applying a binder to the bed, the method including: forming a layer of the powder composition on the bed using the powder feeder; applying binder onto the layer of powder to achieve a binder saturation level ranging from 50-200%; drying the layer of powder for a predetermined time period; and repeating the forming, applying and drying steps until the article is formed. A fifth aspect of the invention provides a method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving powder from the powder feeder, a print head for applying a binder to the bed, the method including: forming a layer of powder composition on the bed using the powder feeder; applying binder to the layer of powder; drying the layer of powder for a predetermined time period; repeating the forming, applying and drying steps until the article is formed; and curing the article in a curing medium for a predetermined curing time period.

BRIEF DESCRIPTION OF THE DRAWINGS

A preferred embodiment of the present invention is hereinafter described by way of example only, with reference to the accompanying drawings, wherein:

Figure 1 is a graph showing the particle size distribution of a powder composition formulated in accordance with an embodiment of the invention.

Figure 2 illustrates images of samples printed using a powder composition formulated in accordance with an embodiment of the invention. Figure 3 is a graph of linear dimensional accuracy results of green cubic specimens printed with (a) Shell/Core = 1 :1 , and (b) Shell/Core = 1 :2 ratios.

Figure 4 is a graph of compressive strength of green cubic specimens printed with (a) Shell/Core = 1 :1 , and (b) Shell/Core = 1 :2 ratios.

Figure 5 is a graph of compressive strengths of a 3D printed article according to an embodiment of the invention cured in“tap water” for 7 and 28 days.

Figure 6 is a graph of compressive strengths of a 3D printed article according to an embodiment of the invention cured in saturated limewater. DESCRIPTION OF EMBODIMENTS

Embodiments of the present invention are illustrated by the following examples by referring to the accompanying description, Figures, and Tables. However, it is to be understood that the scope of the invention is not limited to the embodiments described in the examples, Figures and tables which are merely exemplary in nature to facilitate discussion of the present invention.

Current commercially available powder composition suitable for the construction industry is based on gypsum. One disadvantage of this powder is the low

compressive strength of the material (usually less than 5 MPa). Conventional Portland cement has been considered as the master construction material for its high strength, stability and relatively low cost for over 100 years. However, the setting characteristics of Portland cement limit its use for powder-based 3DCP process. The applicant carried out significant laboratory test work to develop a series of powdered cement-containing compositions according to a first form of the invention that can be used for powder-based 3DCP techniques. In general, the powder composition comprises: (a) an aggregate; (b) an accelerator; (c) a flow control agent and (d) a cementitious material.

A first form of the invention provides a powder composition for 3D printing comprising: (1) a base material (such as a cementitious material);

(2) a set-accelerating material (also known as an accelerator);

(3) a flow control agent; and

(4) an additive material (such as an aggregate).

A table summarizing a selection of powder compositions formulated according to the present invention is presented in Table 1 below. Table 1 : Powder compositions according to the invention.

The“Type GP cement” described in Table 1 above refers to conventional Portland cement conforming to the Australian Standard, AS 3972 general purpose cement. The oxide and mineral compositions of Type GP cement which was used to prepare the compositions Mixes 1-3 of Table 1 are presented in Table 2 below.

Table 2: Exemplary oxide and mineral compositions of conventional Portland cement.

The cementitious material may be Portland cement, and/or calcium sulfoaluminate (CSA) cement. Suitably, the Portland cement is selected from conventional Portland cement, or rapid hardening Portland cement. The accelerator may be one or more of calcium aluminate, calcium sulfate hemihydrate (plaster of Paris) and/or lithium salt. Suitably, the calcium aluminate is one or a mixture of amorphous calcium aluminate and calcium aluminate cement. Suitably, the lithium salt is lithium hydroxide and/or lithium nitrate.

The accelerator may comprise from 1-30wt% of the mass of the cementitious material. The amount of accelerator used may depend on the type of accelerator used. For example, if amorphous calcium aluminate is used as the accelerator, the amount may range from 10-30wt% of the mass of cementitious material. If a lithium salt is used as the accelerator, the amount may range from 1-2% of the mass of the cementitious material. Exemplary specifications of amorphous calcium aluminate (Calumex SC-A®) supplied by Caltra Nederland B.V., Netherlands, which was used to prepare Mix 1 are presented in Table 3 below.

Table 3: Exemplary specifications of amorphous calcium aluminate accelerator.

The composition may include a flow-control agent. Flow-control material may be incorporated into the Portland cement-based powder to increase the flowability and disposability of the powder and reduce the formation of lumps in the powder.

The flow-control agent may have a BET specific surface area ranging from 100-300 m²/g. Suitably, the flow-control agent has a BET specific surface area ranging from 150-250 m²/g. More suitably, the flow-control agent has a BET specific surface area of about 200 m²/g.

Generally, the surface area of the flow-control agent is inversely proportional to flowability. However, flow-control agents that have low surface area are more difficult to handle and typically more expensive.

The flow-control agent may range from 0.5-5wt% of the mass of cementitious material. Suitably, the flow-control agent ranges from 0.5-3wt% of the mass of the cementitious material. More suitably, the flow-control agent ranges from 0.5-2wt% of the mass of the cementitious material. In one example, the amount of flow-control agent ranges from 0.5-1% of the mass of the cementitious material when hydrophilic fumed silica with a BET specific surface area of 2x10⁶ cm²/g (200 m²/g) is used. The flow control agent may be fumed silica. Suitably, the silica is hydrophilic fumed silica.

The flow control agent used to prepare Mixes 1-4 is hydrophilic fumed silica

(AEROSIL® 200) supplied by Evonik Industries AG, Germany. The exemplary specifications for AEROSIL® 200 are presented in Table 4 below. Table 4: Exemplary specifications of hydrophilic fumed silica flow-control agent (AEROSIL® 200).

The additive material may be an aggregate. The aggregate may be silica and/or quartz sand. Suitably, the aggregate is high purity fine silica sand. More suitably, the fine silica sand has a purity ranging from 95-99wt% S1O2. Even more suitably, the fine silica sand has a purity of 99wt% S1O2.

The aggregate may have an average particle size ranging from 100-200 pm.

Suitably, the aggregate has an average particle size ranging from 150-190 pm. More suitably, the aggregate has an average particle size of 184 pm. The inventors discovered that the specified particle size ranges are optimal for adjusting the properties of the powder composition including flowability and wettability.

The aggregate may range from 10% to 300% of the mass of the cementitious material. Suitably, the aggregate ranges from 20-200% of the mass of the cementitious material. The amount of fine silica sand may be used to control the elastic modulus and flowability of the powder which may improve printability.

The inventors discovered that the print quality would be compromised if the aggregate content exceeded the specified ranges. For example, if less than 20% aggregate is used, shrinkage is likely to occur and cause cracks in the printed sample. If more than 200% aggregate is used, this will likely cause insufficient bonding between the cement particles and aggregate, which can result in

disintegration of the sample.

The aggregates used in Mixes 1-4 are a high purity fine silica sand with an average particle size of 184 pm supplied by TGS Industrial Sand Ltd., Australia. The amount of fine silica sand used may range from 20% to 200% of the mass of the cementitious material. It was determined that having aggregate within the described range imparted desirable elastic modulus and flowability properties to the powder. However, it will be appreciated that higher or lower dosages will be appropriate depending on the type of cementitious material. The composition may include reinforcing fibers. The reinforcing fibres may be any one or more of polyethylene (PE) fiber, polyvinyl alcohol (PVA) fiber, calcium carbonate whisker, and/or wollastonite whisker. It will be appreciated that other types of reinforcing fibers may also be suitable.

An example powder composition based on Mix 1 from Table 1 is set out below in Table 5.

Table 5: Example composition based on Mix 1 in Table 1.

It was discovered that the mass percentages presented in Table 1 for this specific example provided desirable setting time and adequate flowability for powder spreading.

A second form of the invention provides a method of forming a powder composition for forming 3D printed article as previously described. The method includes: mixing a cementitious material and an accelerator in a mixer at a speed ranging from 600- 1 ,000 rpm for a time period ranging from 1-10 minutes to form a first mixture; mixing a flow control agent with the first mixture at a speed ranging from 1 ,300-1 ,700 rpm and a time period ranging from 1 ,500 rpm from 1-10 minutes to form a second mixture; and mixing an aggregate with the second mixture at a speed ranging from 1 ,000-3,000 rpm for 1-10 minutes to form the powder composition. The time period for one or more mixing step may be about 5 mins.

The mixing speed for the first mixing step may be about 800 rpm.

The mixing speed for the second mixing step may be about 1 ,500 rpm.

The mixing speed for the third mixing step may be about 2,000 rpm.

Suitably, the method includes: (1) Mixing Type GP cement (Portland cement) and amorphous calcium aluminate in a mixer at 800 rpm for 5 minutes to form a first mixture.

(2) Adding hydrophilic fumed silica to the first mixture and mixing at 1 ,500 rpm for 5 minutes to form a second mixture.

(3) Adding TGS silica sand to the second mixture and mixing at 2,000 rpm for 5 minutes to form the powder composition.

The mixing procedure presented in this specific example was discovered to provide desirable flowability and homogeneity of the powder.

The particle size distribution (PSD) of the example composition based on Mix 1 is presented in Figure 1. Based on the PSD result, the average particle size, Dio, D₅ o and D₉₀ values of the example composition based on Mix 1 are 39.43 pm, 0.69 pm, 17.15 pm and 69.79 pm, respectively.

A third form of the invention provides a method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving a powder composition from the powder feeder, a print head for applying a binder to the bed, the method including: forming a layer of the powder composition as previously described on the bed using the powder feeder; applying binder to the layer of powder; drying the layer of powder for a predetermined time period; and repeating the forming, applying and drying steps until the article is formed. The step of forming a layer of the powder composition may involve forming a layer of about 0.090-0.150 mm thickness. More suitably, the step involves forming a layer of about 0.100-0.120 mm thickness. Even more suitably, the step involves forming a layer of about 0.102 mm.

The step of forming a layer of the powder composition may involve using a roller. The step of applying binder to the layer of powder may involve adding droplets of binder to the layer of powder. Suitably, the step of applying binder to the layer of powder may include controlling the size of the droplets.

The binder may have a viscosity similar to water.

The predetermined time period may range from 200 to 1 ,500 ms. Suitably, the predetermined time period ranges from 300-1 ,000 ms.

The predetermined time period may be dependent on one or more of (i) the dosage of accelerator, (ii) the pore structure within the powder bed, and (iii) the level of binder saturation.

Binder saturation is defined as the ratio between the volume of deposited liquid binder and the volume of pores in the powder bed. Two sub-variables of the binder saturation are employed in the powder-based 3D printers, namely shell and core. The shell refers to the region comprising the edges of the sample and parts of the interior area within the edges of the sample. The core refers to the remaining interior areas within the edges of the sample. High binder saturation results in bleeding of the binder into the surrounding powder while low binder saturation may result in the printed article having weak green strength due to the poor bonding between particles of the powder.

The step of applying binder to the layer of powder may involve applying binder to provide a binder saturation level ranging from 50-200%. Suitably, the binder saturation level ranges from 90-180%. In one embodiment, the binder saturation level is 100%. In a second embodiment, the binder saturation level is 135%. In a third embodiment, the binder saturation level is 170%.

The step of applying binder to the layer of powder may involve applying binder to provide a Shell/Core ratio ranging from 3: 1 to 1 :3. Suitably, the shell/core ratio ranges from 2: 1 to 1 :2. In one embodiment, the shell/core ratio is 1 :1. In a second embodiment, the shell/core ratio is 1 :2.

The method may include a step of curing the article in a curing medium for a predetermined curing time period. Suitably, the predetermined curing time period ranges from 5-9 days. More suitably, the predetermined curing time period is about 7 days. The curing medium may be an aqueous solution. Suitably the curing medium is either tap water or limewater. The limewater may be saturated limewater.

EXAMPLE 1

The 3D printed specimens printed using the example powder composition based on Mix 1 are presented in Figure 2. After the step of removing unbound powder, the 3D printed cube samples with the dimensions of 20 x 20 x 20 mm (shown in Figure 2(c)) were divided into two groups. For the first group denoted as“green” samples, no further post-processing procedures is undertaken. For the second group denoted as“post-processed” samples, the printed cubes are immersed in tap water and kept at ambient temperature (23 ± 3 °C) for 28 days. At the end of the curing period, the ambient temperature cured samples are taken out from the curing medium and kept undisturbed until become dry. The compressive strengths in both X-direction (i.e. the binder jetting direction) and Z-direction (i.e. layer stacking direction) are measured under load control at the rate of 0.33 MPa/s. A population of 10 samples for each testing direction was used. The compressive strength test results are summarized in Table 6.

Table 6: Test results for 3D printed cube samples made from Mix 1 composition.

Linear Dimensional Accuracy

One of the important factors that define printability is printing accuracy. Figure 3 is a graph of the results of the linear dimensional accuracy of green cubic samples.

According to Fig. 3, the mean error values of the samples printed with saturation levels of 135% and 170% in all directions were always greater than zero. This indicates that the measured dimensions of printed samples in all directions were more than those of the digital model. This pattern is true regardless of the shell to core ratio. With regards to the samples printed with a saturation level of 100%, the measured dimensions of the samples printed with Shell/Core of 1 : 1 were less than the digital model, whereas the measured dimensions of the samples printed with Shell/Core of 1 :2 were more than the digital model. It should be noted that an anisotropic phenomenon was observed regarding the linear dimensional accuracy of the green samples depending on the testing directions. For all binder saturation levels, the Z-direction had the highest mean error values. In other words, regardless of the binder saturation level, the Z-direction had always the lowest linear dimensional accuracy. This phenomenon is more pronounced in the samples printed with Shell/Core of 1 :2. This anisotropic phenomenon might be associated with the different rates of binder penetrating in vertical (Z) direction and spreading in lateral (X and Y) directions. On the other hand, for all binder saturation levels, the X-direction had the lowest mean error values, thereby the highest linear dimensional accuracy. The inventors consider that this may be due to the X-direction (i.e. the binder jetting direction) not being affected by the powder spreading that takes place in Y-direction.

As can be seen in Fig. 3, the increase in the binder saturation level considerably increased the mean error values in all directions. For example, for the green samples printed with Shell/Core of 1 :1 the mean error value in Z-direction significantly increased from -0.07 ± 0.08 mm in the case of S100C100 to 0.24 ± 0.17 mm in the case of S170C170. In other words, the increase in the binder saturation level significantly reduced the linear dimensional accuracy of the green samples in all directions. According to Fig. 3, the change in the Shell/Core ratio from 1 :1 to 1 :2 significantly increased the mean error values in all directions. This is true regardless of the binder saturation level. For instance, for the samples printed with Shell/Core of 1:1 the mean error value in Z-direction significantly increased from 0.13 ± 0.23 mm in the case of S170C170 to 0.57 ± 0.18 mm in the case of S170C340 with a Shell/Core of 1 :2. In summary, it can be concluded that the increase in the binder saturation level and/or the shell to core ratio significantly reduced the linear dimensional accuracy of the green samples printed using the developed Portland cement-based powder. This might be explained by the bleeding mechanism, since at higher binder saturation level the excess binder spreads outside the edges of the printed sample, which results in a reduction of the linear dimensional accuracy.

Compressive Strength

Figure 4 shows the compressive strength of green samples printed using the developed Portland cement-based powder in both X-direction and Z-direction.

As is shown in Fig. 4, an anisotropic phenomenon was observed regarding the compressive strength of the green samples depending on the loading directions.

Regardless of the saturation level and shell to core ratio, the compressive strength was always higher in X-direction than in Z-direction. The anisotropy effect was more pronounced in the samples printed with Shell/Core of 1 :2 as compared to Shell/Core of 1 :1. For example, the compressive strength of the green samples printed with S100C100 in X-direction was 14% higher than that in Z-direction. However, the compressive strength of the green samples printed with S100C200 in X-direction was 34% higher than that in Z-direction. It is believed that this anisotropic phenomenon might be related to the preferential orientation of the powder particles during the powder spreading process.

According to Fig. 4, the change in the Shell/Core ratio from 1 :1 to 1 :2 significantly increased the compressive strength of the green samples. This is true regardless of the testing direction and binder saturation level. It is interesting to note that the rate of increase in the compressive strength of the green samples in X-direction was higher than in Z-direction. For instance, in X-direction the compressive strength of

S100C100 samples was 4.9 times lower than that of the S100C200 samples.

However, the corresponding value in Z-direction was 4.1 times. The green samples printed with Shell/Core of 1 :1 exhibited relatively low compressive strength ranging from 0.7 to 1.5 MPa, depending on the testing direction and saturation level.

However, it should be noted that the strength of all green samples printed with Shell/Core of 1 : 1 was already more than enough for the de-powdering process. In both directions, the increase in the binder saturation level significantly increased the compressive strengths. However, the rate of increase in the compressive strength of the green samples with Shell/Core of 1 :2 was higher than that with Shell/Core of 1 : 1. For instance, for Shell/Core of 1 :1 and in X-direction the compressive strength of S170C170 samples was 1.9 times higher than that of S100C100 samples. However, the corresponding value for the Shell/Core of 1 :2 is 2.15 times.

The inferior compressive strength of the green samples with lower binder saturation level and shell to core ratio is probably due to incomplete hydration process, which results in a weak bond between powder particles. Higher binder saturation level and core saturation use a higher volume of binder during the printing process, resulting in superior bonding between the powder particles.

EXAMPLE 2

In a separate set of experiments, the same 3D printed specimens printed using the Mix 1 composition was produced. After the step of removing unbound powder, the printed samples were subjected to“tap water curing” and“saturated limewater curing”. The compressive strengths of cured samples were measured at 7 and 28 days. However, the compressive strengths of the green samples were tested 2 h after the end of the printing process (i.e., after the de-powdering process). The compressive strengths in both X-direction (i.e., the binder jetting direction) and Z- direction (i.e., layer stacking direction) were measured under load control at the rate of 0.33 MPa/s.

Tap water curing

Figure 5 shows the 7-day and 28-day compressive strengths of the tap water cured samples. The compressive strengths of the green samples are also presented in this figure. The error bars in the presented results are based on 95% confidence level and the numerical values in the middle of the bars are the mean strengths. As can be seen in Figure 5, the compressive strength of the post-processed printed samples cured in tap water was significantly higher than that of the green samples. The 7-day compressive strength of the tap water cured printed samples was about

2.1 times higher than the compressive strength of the green samples. In addition, according to Figure 5, the longer the curing time, the higher the strength gain. The

28-day compressive strength of the tap water cured printed samples was more than

2.2 times higher than the 7-day compressive strength. This is true regardless of the loading directions. This significant increase in strength is due to the continued cement hydration of the un-reacted cement particles in the presence of tap water. As shown in Figure 5, an anisotropic phenomenon was observed in the compressive strength of the printed samples depending on the loading directions. Regardless of the age of testing, the mean compressive strength was always higher in the X- direction than in the Z-direction. The anisotropy in compressive strength might be related to the bond strength between layers. In the green samples, it is believed that the water content significantly oscillates in accordance with a higher water content in the top region of the layer and a significantly lower content in the bottom region.

Thus, the intralayer water gradient will cause a weak bond between layers.

However, as can be seen in Figure 5, the Wfc-z (ratio of compressive strength in X- direction to the compressive strength in Z-direction) decreased from 1.12 to 1.09 with the increase of the curing time, which implies that the degree of anisotropy in compressive strength was reduced with the increase of curing time. The reason for this can be explained by the improved inter-layer bond strength resulting from the continued cement hydration of the un-reacted cement particles in the curing process.

Saturated limewater curing Figure 6 shows the compressive strength of the green samples and the 7-day and 28-day compressive strengths of samples cured in saturated limewater. The error bars in the presented results are based on 95% confidence level and the numerical values in the middle of the bars are the mean strengths.

According to Figure 6, the 7-day compressive strength of the post-processed printed samples cured in saturated limewater was about 2.6 times higher than the compressive strength of the green samples. In addition, the 28-day compressive strength of the saturated limewater cured printed samples was about 2.1 times higher than the 7-day compressive strength. This is true regardless of the loading directions.

Comparison between Figures 5 and 6 shows that all printed samples cured in saturated limewater generally acquired higher strength than those cured in tap water. This is also true regardless of the at all curing time and loading direction. Regardless of the loading direction, the 7-day and 28-day compressive strengths of the saturated limewater cured samples were 26% and 17%, respectively higher than the 7-day and 28-day compressive strengths of the tap water cured samples. Since the

concentration of Ca²⁺ in the saturated limewater is high and CH is quite soluble in cementitious materials, as cement hydration proceeded with time, less dissolution and leaching of CH occurred in the saturated limewater curing medium than the tap water, therefore results in higher compressive strength.

The compressive strength of the saturated limewater cured samples also exhibited similar anisotropic behavior, depending on the testing direction. The mean compressive strength was always higher in the X-direction than in the Z-direction, regardless of the age of testing. Similar to the printed samples cured in tap water, the Wfc-z also decreased with the increase of the curing time, indicating that the degree of anisotropy in compressive strength was reduced with the increase of the curing time.

The post-processed 3D printed articles exhibited a 28-day compressive strength of 23.1-29.4 MPa, depending on the type of curing medium and testing direction, which is 4.2 to 6.0 times higher than the compressive strength of the green samples (4.9- 5.5 MPa). The following observations were drawn from the experimental data:

1. The increase in the binder saturation level and/or the core saturation significantly reduced the linear dimensional accuracy of the green samples. This is true regardless of the testing direction. This is probably because the excess binder at higher binder saturation spreads outside the edges of the printed sample, resulting in reduction of the linear dimensional accuracy.

2. The linear dimensional accuracy of the green samples exhibited an anisotropic behaviour depending on the testing direction. It was observed that the Z- direction always had the lowest linear dimensional accuracy. The inventors believe this may be the result of the different rates of binder penetrating in the vertical (Z) direction and spreading in the lateral (X and Y) directions. On the other hand, it was observed that the X-direction always had the highest linear dimensional accuracy, which the inventors believe is due to the X-direction (i.e. , the binder jetting direction) not being affected by the powder spreading that takes place in Y-direction. This orthotropic behaviour was true regardless of the binder saturation level, but more pronounced in the samples printed with Shell/Core of 1 :2 than 1 :1.

3. The compressive strength of the green samples also exhibited an orthotropic behaviour depending on the testing direction. Regardless of the saturation level and shell to core ratio, the compressive strength in X-direction was always higher than in Z-direction, which the inventors believe may be related to the preferential orientation of the powder particles during the powder spreading process. This orthotropic behaviour was more pronounced in the samples printed with Shell/Core of 1 :2 than 1 : 1. 4. The change in the Shell/Core ratio from 1 :1 to 1 :2 significantly increased the compressive strength of the green samples. This is true regardless of the testing direction and binder saturation level.

5. In both X and Z directions, the increase in the binder saturation level significantly increased the compressive strengths. However, the rate of increase in the compressive strength of the green samples with Shell/Core ratio of 1 :2 was higher than that with Shell/Core ratio of 1 : 1.

6. The inferior compressive strength of the green samples printed with lower binder saturation levels and Shell/Core ratios is probably due to incomplete hydration process due to insufficient amount of binder, which in turn results in a weak bond between the powder particles.

7. Tap water and saturated limewater are both effective post-processing curing mediums to enhance the strength of 3D printed articles. The significantly higher strength of the post-processed printed samples is due to the continued cement hydration of the un-reacted cement particles in the presence of either tap water or saturated limewater.

8. Among the two curing mediums investigated in this study, saturated limewater was more effective in enhancing the strength of 3D printed article. This is due to the less CH leaching and dissolution in the saturated limewater which favored the continued cement hydration process.

9. An anisotropic phenomenon was observed in the compressive strength of the 3D printed articles depending on the loading directions. The compressive strength in X-direction (i.e., the binder jetting direction) was always higher than in Z-direction (i.e., layer stacking direction). This is true regardless of the curing medium and age of testing.

10. The fc-x/fc-z (i.e., the ratio of compressive strength in X-direction to the compressive strength in Z-direction) decreased with the increase of curing time. Therefore, it is concluded that the degree of anisotropy in compressive strength was reduced with the increase of curing time. This is true regardless of the curing medium. This reduction in the degree of anisotropy is attributed to the improved inter- layer bond strength resulted from the continued cement hydration of the un-reacted cement particles in the curing process.

It is to be understood that, if any prior art publication is referred to herein, such reference does not constitute an admission that the publication forms a part of the common general knowledge in the art, in Australia or any other country. In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word“comprise” or variations such as“comprises” or“comprising” is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.

EXPERIMENTAL

Materials

A conventional Portland cement conforming to the Australian Standard, AS 3972 general purpose (Type GP) cement was used in this study. The percentages of C3S, C2S, C3A and C4AF as the main constituents of Portland cement were 57.59%, 14.87%, 4.10% and 13.94%, respectively. A small amount of accelerator was used to reduce the setting time of Portland cement, making it suitable for the powder-based 3DCP process. A high purity fine silica sand with an average particle size of 184 pm supplied by TGS Industrial Sand Ltd., Australia was also used in this study to function as an inert filler in the Portland cement-based powder to improve its printability. The raw materials were thoroughly dry mixed for 10 min in an Eirich mixer to achieve a homogenous mixture (visually assessed). The particle size distribution (PSD) of the developed Portland cement-based powder is given in Fig. 1. The PSD analysis result showed the average particle size, D10, D50 and D90 values of the powder were 39.43 pm, 0.69 pm, 17.15 pm and 69.79 pm, respectively.

Methods

3D Printing Process. A commercial powder-based 3D printer (Zprinter® 150) manufactured by Z-Corp, USA was used. A Cl LAS particle size analyzer model 1190 was used to obtain the PSD.

The samples were printed using a commercially available powder-based 3D printer (Zprinter® 150, Z-Corp, USA) with an HP11 print head (C4810A). The Zprinter® 150 3D printer has a specified resolution of 300^450 dpi and a build speed of 2-4 layers per minute. The HP11 print head (C4810A) has 304 nozzles with a drop size of 2.5 pL for each nozzle. A commercial clear binder solution (Zb® 63) with a similar viscosity to pure water supplied by the printer’s manufacture was used.

The mechanism of controlling the binder solution is a non-continuous approach called drop-on-demand (DoD) technique.

Linear Dimensional Accuracy. A digital Vernier calliper with an accuracy of up to 0.01 mm was used to measure the dimensions of green cubic samples in three directions, namely X-direction (the direction of binder jetting), Y-direction (the direction of powder layer spreading) and Z-direction (the direction of layer stacking). The linear dimensional error was calculated based on the following equation:

Equation L_aciuai L_nominai

Where L_actuai is the measured length, whereas L_n0minai is the length of the digital model. A population of 10 samples for each testing direction was used. For each sample, three measurements were taken for each testing direction and the mean error values were calculated to assess the linear dimensional accuracy.

Compressive Strength. The compressive strengths of green cubic samples in both X- direction and Z-direction were measured under load control at the rate of 0.33 MPa/s. A population of 10 samples for each testing direction was used.

Post-processing and testing methods. After the de-powdering process, the printed samples were divided into two groups denoted as“tap water curing” and“saturated limewater curing”. The saturated limewater was prepared by adding 2 grams of

Ca(OH)₂ powder in 1 liter of tap water and slightly stirred. For the“tap water curing” group, the printed cubes were immersed in a sealed container containing tap water, while for the“saturated limewater curing”, the printed cubes were immersed in a sealed container containing saturated limewater. Both containers were kept at ambient temperature (23 ± 3 °C) for 7 and 28 days. At the end of the curing period, the cured samples were taken out from the curing mediums and kept undisturbed until become dry. The compressive strengths of cured samples were measured at 7 and 28 days. Flowever, the compressive strengths of the green samples were tested 2 h after the end of the printing process (i.e., after the de-powdering process). The compressive strengths in both X-direction (i.e., the binder jetting direction) and Z- direction (i.e., layer stacking direction) were measured under load control at the rate of 0.33 MPa/s. A population of 10 samples for each testing direction was used.

Claims

THE CLAIMS:

1. A powder composition for forming 3D printed article, the composition comprises:

an aggregate;

a cementitious material selected from Portland cement and/or Calcium

Sulfoaluminate (CSA) cement to bind the aggregate when mixed with a binder to form a 3D printed article; and

an accelerator selected from calcium aluminate, calcium sulfate hemihydrate (plaster of Paris) or lithium salt to accelerate setting of the article formed; and

a flow control agent to control flowability of the powder.

2. The powder composition according to claim 1 , comprising Portland cement, amorphous calcium aluminate, hydrophilic fumed silica, and an aggregate.

3. The powder composition according to claim 1 , comprising Portland cement, calcium aluminate cement, calcium sulfate hemihydrate, hydrophilic fumed silica and an aggregate.

4. The powder composition according to claim 1 , comprising Portland cement, lithium salt, hydrophilic fumed silica and an aggregate.

5. The powder composition according to claim 1 , comprising hardening Portland cement, hydrophilic fumed silica and an aggregate.

6. The powder composition according to claim 1 , wherein the accelerator comprises from 1-30wt% of the mass of the cementitious material.

7. The powder composition according to either claim 1 or 2, wherein the flow- control agent has a BET specific surface area ranging from 100-300 m²/g.

8. The powder composition according to any one of the preceding claims, wherein the flow-control agent ranges from 0.5-5wt% of the mass of cementitious material.

9. The powder composition according to any one of the preceding claims, wherein the aggregate has an average particle size ranging from 100-200 pm.

10. The powder composition according to any one of the preceding claims, wherein the aggregate has range from 10% to 300% of the mass of the cementitious material.

11. The powder composition according to any one of the preceding claims, including reinforcing fibres selected from any one or more of polyethylene (PE) fiber, polyvinyl alcohol (PVA) fiber, calcium carbonate whisker, and/or wollastonite whisker.

12. A method of forming a powder composition for forming 3D printed article according to any one of the preceding claims, the method including:

mixing a cementitious material and an accelerator in a mixer to form a first mixture; mixing a flow control agent with the first mixture to form a second mixture; and mixing an aggregate with the second mixture to form the powder composition.

13. The method according to claim 12, wherein the time period for one or more mixing steps is about 5 mins.

14. The method according to either claim 12 or 13, wherein the mixing speed for the first mixing step is about 800 rpm.

15. The method according to any one of claims 12 to 14, wherein the mixing speed for the second mixing step is about 1 ,500 rpm.

16. The method according to any one of claims 12 to 15, wherein the mixing speed for the third mixing step is about 2,000 rpm.

17. The method according to any one of claims 12 to 16, including:

mixing Portland cement and amorphous calcium aluminate in a mixer at 800 rpm for 5 minutes to form a first mixture;

adding hydrophilic fumed silica to the first mixture and mixing at 1 ,500 rpm for 5 minutes to form a second mixture; and

adding silica sand to the second mixture and mixing at 2,000 rpm for 5 minutes to form the powder composition.

18. A method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving a powder composition from the powder feeder, a print head for applying a binder to the bed, the method including:

forming a layer of the powder composition according to any one of claims 1 to 11 on the bed using the powder feeder;

applying binder to the layer of powder;

drying the layer of powder for a predetermined time period; and

repeating the forming, applying and drying steps until the article is formed.

19. A method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving powder from the powder feeder, a print head for applying a binder to the bed, the method including:

forming a layer of the powder composition on the bed using the powder feeder; applying binder onto the layer of powder to achieve a binder saturation level ranging from 50-200%;

drying the layer of powder for a predetermined time period; and

repeating the forming, applying and drying steps until the article is formed.

20. A method of forming a 3D printed article, using a 3D printer comprising a powder feeder, a bed for receiving powder from the powder feeder, a print head for applying a binder to the bed, the method including:

forming a layer of powder composition on the bed using the powder feeder;

applying binder to the layer of powder;

drying the layer of powder for a predetermined time period;

repeating the forming, applying and drying steps until the article is formed; and curing the article in a curing medium for a predetermined curing time period.

21. The method according to any one of claims 18-20, wherein the step of forming a layer of the powder composition involves forming a layer of about 0.090- 0.150 mm thickness.

22. The method according to any one of claims 18-21 , wherein the step of applying binder to the layer of powder involves adding droplets of binder to the layer of powder.

23. The method according to any one of claims 18-22, wherein the step of applying binder to the layer of powder achieves a Shell/Core ratio ranging from 3:1 to 1 :3.

24. The method according to either claims 18 or 20, wherein the step of applying binder to the layer of powder achieves a binder saturation level ranging from 50- 200%.

25. The method according to either claims 18 or 19, including a step of curing the article in a curing medium for a predetermined curing time period.

26. The method according to either claim 20 or 25, wherein the curing medium is either water or limewater.

27. The method according to any one of claims 20, 25 or 26, wherein the curing time period ranges from 5-9 days.